[Mauiusers] Re: Problems with Maui/SGE - jobs won't run

Marko marko at doc.ic.ac.uk
Tue Nov 22 02:21:25 MST 2005


Hi again,

It seems that something is wrong with the configuration as
Maui doesn't detect any resources/workload:
   > 11/17 19:53:16 WARNING:  no resources detected
   > 11/17 19:53:16 WARNING:  no workload detected

I've tried installing and configuring SGE/MAUI again, but
unfortunately I get the same result. Could somebody please
comment on my email below.

Thanks in advance,
Marko


Marko wrote:
> Hi all,
> 
> I've have a simple two node test cluster with SGE (v5.3p6)
> sitting on top of it. SGE was built from source and
> was working fine, i.e. jobs were executed.
> 
> I then installed Maui (v3.2.6p13) and tried to integrate it
> with SGE following documentation available on:
> http://www.clusterresources.com/products/maui/docs/sgeintegration.shtml
> 
> How do I now submit a job? Is it the same qsub command as
> before? For example, 'qsub test.sh' with test.sh below
>   #!/bin/sh
>   #$ -S /bin/sh
>   #$ -cwd
>   /bin/hostname --long
> worked fine prior to installing Maui (and stopping SGE
> scheduler in the process), but now job just stays in the
> queue.
> 
> I also tried to submit the same job to a parallel environment
> created during Maui/SGE configuration process:
>   qsub -pe default 1 test.sh
> 
> I'm probably missing something trivial, but I've been trying
> for many hours to solve this issue.
> 
> Thanks in advance,
> Marko
> 
> P.S.
> If it helps, below is a part from maui.log:
> 
> 11/17 19:53:16 ServerProcessRequests()
> 11/17 19:53:16 INFO:     not rolling logs (597316 < 10000000)
> 11/17 19:53:16 MResAdjust(NULL,0,0)
> 11/17 19:53:16 MStatInitializeActiveSysUsage()
> 11/17 19:53:16 MStatClearUsage([NONE],Active)
> 11/17 19:53:16 ServerUpdate()
> 11/17 19:53:16 MSysUpdateTime()
> 11/17 19:53:16 INFO:     starting iteration 247
> 11/17 19:53:16 MRMGetInfo()
> 11/17 19:53:16 MClusterClearUsage()
> 11/17 19:53:16 MRMClusterQuery()
> 11/17 19:53:16 WARNING:  no resources detected
> 11/17 19:53:16 MRMWorkloadQuery()
> 11/17 19:53:16 WARNING:  no workload detected
> 11/17 19:53:16 MStatClearUsage(node,Active)
> 11/17 19:53:16 MClusterUpdateNodeState()
> 11/17 19:53:16 MQueueSelectAllJobs(Q,HARD,ALL,JIList,DP,Msg)
> 11/17 19:53:16 MQueueSelectAllJobs(Q,SOFT,ALL,JIList,DP,Msg)
> 11/17 19:53:16 
> MQueueSelectJobs(SrcQ,DstQ,HARD,5120,4096,2140000000,EVERY,FReason,FALSE)
> 11/17 19:53:16 
> MQueueSelectJobs(SrcQ,DstQ,SOFT,5120,4096,2140000000,EVERY,FReason,TRUE)
> 11/17 19:53:16 
> MQueueSelectJobs(SrcQ,DstQ,HARD,5120,4096,2140000000,EVERY,FReason,TRUE)
> 11/17 19:53:16 
> MQueueSelectJobs(SrcQ,DstQ,SOFT,5120,4096,2140000000,EVERY,FReason,TRUE)
> 11/17 19:53:16 MSchedUpdateStats()
> 11/17 19:53:16 INFO: iteration: 247  scheduling time: 0.000 seconds
> 11/17 19:53:16 MResUpdateStats()
> 11/17 19:53:16 INFO: current util[247]: 0/0 (0.00%)  PH: 0.00%  active 
> jobs: 0 of 0 (completed: 0)
> 11/17 19:53:16 MQueueCheckStatus()
> 11/17 19:53:16 MNodeCheckStatus()
> 11/17 19:53:16 MUClearChild(PID)
> 11/17 19:53:16 INFO: scheduling complete.  sleeping 30 seconds



More information about the mauiusers mailing list